PSY201H1 Lecture Notes - Lecture 4: Pearson Product-Moment Correlation Coefficient, Missing Data, Venn Diagram
Document Summary
Individual points plotted displayed relationship between x and y. Line of best fit drawn in where largest density of points are/most points are touching/close to line. Perfect relationship when all points fall along line; extremely rare: but in this case when you know x, you can find y and vice versa7. Help us choose the right analysis: linear = pearson correlation coefficient. If not, see text: non-linear = special treatment (beyond this course, check for outliers. Convert all your x scores to z-scores using the sd and mean of x. Convert all your y scores to z-scores using the sd and mean of y. Multiply your matching zx and zy scores. Then divide by the number of pairs of scores you have minus 1. Reminds us that the correlation is standardized: regardless of scale of variables. but what about the other 79. 75%: error in prediction, other stuff. Check your assumptions: normality, skew, kurtosis, missing data, the dreaded outliers.
